Skip to content

See docs/DISCLAIMER_SNIPPET.md

Meta‑Agentic α‑AGI 👁️✨ Demo v2 – Production‑Grade v0.1.0

preview

Launch Demo

Current runnable path — 1.14.0

Mode: Synthetic evaluation. Runs provider-driven code proposals, synthetic fitness and SQLite lineage.

Prerequisites: Python 3.11–3.13; installed project dependencies.

After installation:

python -m alpha_factory_v1.demos check meta_agentic_agi_v2
python -m alpha_factory_v1.demos run meta_agentic_agi_v2

The catalog command uses bundled inputs and explicit offline defaults.

Expected result: Three generations saved to lineage.sqlite; repeat runs append safely.

Scope: The mock provider and pseudo-accuracy are explicitly synthetic, not measured task performance.

The catalog explains installation, stopping, backups and recovery. Browser charts for legacy demos are labeled sample replays. Original research narratives and advanced scripts below are preserved; they do not expand the tested scope stated here.

This repository is a conceptual research prototype. References to "AGI" and "superintelligence" describe aspirational goals and do not indicate the presence of a real general intelligence. Use at your own risk. Nothing herein constitutes financial advice. MontrealAI and the maintainers accept no liability for losses incurred from using this software. Each demo package exposes its own __version__ constant. The value marks the revision of that demo only and does not reflect the overall Alpha‑Factory release version.

Identical to v1 plus a statistical-physics wrapper that logs and minimises Gibbs / variational free-energy for each candidate agent during the evolutionary search.

Metric toggle: configs/default.yml → physics_metric: free_energy
Implementation: core/physics/gibbs.py (≈25 LoC) & 4-line hook in scorer.py.

Official definition – Meta-Agentic (adj.)
Describes an agent whose primary role is to create, select, evaluate, or re‑configure other agents and the rules governing their interactions, thereby exercising second‑order agency over a population of first‑order agents.

The term was pioneered by Vincent Boucher, President of MONTREAL.AI.

%% 𝗚𝗿𝗮𝗻𝗱 𝗦𝘆𝗻𝗮𝗽𝘀𝗲 𝗚𝗿𝗮𝗽𝗵 – Meta-Agentic α-AGI (v2 with variational free-energy)
graph LR
  %% ─────────────────────  Class styles
  classDef meta      fill:#6425ff,stroke:#eee,color:#fff
  classDef layer     fill:#1e1e2e,stroke:#ddd,color:#fff
  classDef agent     fill:#0f9d58,stroke:#fff,color:#fff
  classDef tool      fill:#fbbc05,stroke:#000,color:#000
  classDef physics   fill:#ff6d00,stroke:#000,color:#fff

  %% ─────────────────────  High-level layers
  A0["🧠 Meta-Programmer"]:::meta
  A1["📈 Evolution Archive"]:::layer
  A2["⚖️ Multi-Objective Scorer"]:::layer
  Aφ["♾️ Free-Energy Monitor"]:::physics
  A3["🧩 Agent Population"]:::layer

  %% ─────────────────────  First-order agents
  subgraph " "
    direction TB
    D1["🔍 Researcher"]:::agent
    D2["👷 Builder"]:::agent
    D3["🧪 Evaluator"]:::agent
    D4["⚙️ Auto-Tuner"]:::agent
    D5["🛡 Guardian"]:::agent
  end

  %% ─────────────────────  Foundation-model providers / tools
  subgraph " "
    direction TB
    T1["GPT-4o"]:::tool
    T2["Claude-3"]:::tool
    T3["Llama-3 ∞"]:::tool
  end

  %% ─────────────────────  Core data/value loop
  subgraph " "
    direction LR
    V1["🌐 Industry Data Streams"]
    V2["💎 Extracted Alpha"]
    V3["🚀 Deployed Solutions"]
  end

  %% ─────────────────────  Arrows
  A0 -->|generate| A3
  A3 -->|select| A2
  A2 -->|rank| A1
  A1 -- feedback --> A0

  %% Free-energy feedback
  A3 -.state logits.-> Aφ
  Aφ -->|F metric| A2
  Aφ -- entropy gradient --> A0

  %% Providers
  D1 -.uses.-> T1
  D2 -.uses.-> T3
  D3 -.uses.-> T2
  D4 -.uses.-> T3
  D5 -.uses.-> T1

  %% Value extraction
  A3 -->|iterate| V1
  V1 -->|signals| D1
  D2 --> V2
  V2 --> D3
  D4 --> V3
  D5 -.audit.-> V3

Elevating Alpha‑Factory v1 into a self‑improving, cross‑industry “Alpha Factory” that systematically
Out‑Learn · Out‑Think · Out‑Design · Out‑Strategize · Out‑Execute — without coupling to a single vendor or model.
Inspired by and extending the Meta‑Agent Search paradigm from Hu et al. (ICLR 2025).

📌 Purpose & Positioning

This demo operationalises the Automated Design of Agentic Systems (ADAS) paradigm and layers:

  • True multi‑objective search (accuracy, cost, latency, risk, carbon)
  • Open‑weights or API‑based FM back‑ends (OpenAI, Anthropic, Mistral .gguf …)
  • Automated provenance & lineage visualisation
  • Antifragile, regulator‑ready safeguards

into the existing Alpha‑Factory v1 (multi‑agent AGENTIC α‑AGI) pipeline.

1 Quick‑start 🏁

# 1️⃣ Clone & enter demo
git clone https://github.com/MontrealAI/AGI-Alpha-Agent-v0.git
cd AGI-Alpha-Agent-v0/alpha_factory_v1/demos/meta_agentic_agi_v2

# 2️⃣ Environment (CPU‑only default)
micromamba create -n metaagi python=3.11 -y
micromamba activate metaagi
pip install -r requirements.txt        # ≤ 40 MiB wheels

# 3️⃣ Run – zero‑API mode (pulls a gguf via Ollama)
python meta_agentic_agi_demo_v2.py --provider mock:echo       # offline demo
#   …or real weights
python meta_agentic_agi_demo_v2.py --provider mistral:7b-instruct.gguf

#   …or point to any provider
OPENAI_API_KEY=sk‑… python meta_agentic_agi_demo_v2.py --provider openai:gpt-4o

# 4️⃣ Launch the lineage UI
streamlit run ui/lineage_app.py

No GPU? llama‑cpp‑python auto‑selects 4‑bit quantisation < 6 GB RAM.

🎓 Colab notebook

Open In Colab

Spin up the demo end‑to‑end without installing anything. Works offline using open‑weights or with your API keys. The notebook now previews the latest lineage entries after the search loop so you can inspect results directly in Colab.

2 Folder Structure 📁

meta_agentic_agi_v2/
├── core/                # provider‑agnostic primitives
│   ├── fm.py            # unified FM wrapper
│   ├── prompts.py       # reusable prompt fragments
│   └── tools.py         # exec sandbox, RAG, vector store
├── meta_agentic_search/ # ⬅ evolutionary loop
│   ├── archive.py       # stepping‑stone JSONL log
│   ├── search.py        # NSGA‑II + Reflexion
│   └── scorer.py        # multi‑objective metrics
├── agents/
│   ├── agent_base.py    # runtime interface
│   └── seeds.py         # bootstrap population
├── ui/
│   ├── lineage_app.py   # Streamlit dashboard
│   └── assets/
├── configs/
│   └── default.yml      # editable in‑UI
└── meta_agentic_agi_demo_v2.py

3 High‑Level Architecture 🔍

graph TD
  subgraph "Meta Agent Search Loop"
    MGPT["Meta LLM Programmer"]
    Candidate["Candidate Agent<br/>Python fn"]
    Evaluator["Sandboxed Evaluator"]
    Archive["Archive<br/>(Pareto + Novelty)"]
    MGPT -->|generates| Candidate
    Candidate --> Evaluator
    Evaluator -->|scores| Archive
    Archive -->|context & feedback| MGPT
  end
  UI["Streamlit Lineage UI"] <-->|stream
lineage| Archive

flowchart LR
  AFV1["Alpha‑Factory v1 Core"]
  MAA["Meta‑Agentic Layer"]
  Providers["FM Providers<br/>(OpenAI / Anthropic / llama‑cpp)"]
  Dataset["Domain Datasets"]
  UI2["Lineage UI"]
  AFV1 --> MAA
  MAA --> Providers
  MAA --> Dataset
  Archive -.-> UI2

4 Provider Abstraction ➡️ open‑weights 🏋️‍♀️

configs/default.yml (excerpt):

provider: mistral:7b-instruct.gguf   # any ollama / llama.cpp id
context_length: 8192
rate_limit_tps: 4
retry_backoff: 2

Change provider to:

Value Notes
openai:gpt-4o needs OPENAI_API_KEY
anthropic:claude-3-sonnet needs ANTHROPIC_API_KEY
mistral:7b-instruct.gguf default local model
mock:echo offline stub, for tests
---

Objective vector = [accuracy, cost, latency, hallucination‑risk, carbon]

  • NSGA‑II elitist selection
  • Behaviour descriptor = SHA‑256 of candidate AST
  • Optional human‑in‑the‑loop thumbs up/down (UI)

6 Security & Antifragility 🛡

  • Firejail --seccomp + 512 MiB mem‑cgroup sandbox
  • Static analysis (bandit) + dynamic taint tracking
  • Live watchdog kills rogue processes > 30 s CPU
  • Chaos‑tests inject tool failures; reward graceful degradation

7 Extending 🛠

  1. New dataset – drop my.pkl into data/, flag --dataset my.
  2. New metric – subclass scorer.BaseMetric, list in configs/default.yml.
  3. New tool – add core/tools/foo.py exposing __call__(self, query).

8 Roadmap 🗺

  • ☐ Hierarchical meta‑meta search
  • ☐ GPU batch infer (Flash‑infer v3)
  • ☐ Offline RL fine‑tune search policy with lineage replay

9 References 📚

  • S. Hu et al. “Automated Design of Agentic Systems” ICLR 2025
  • OpenAI “A Practical Guide to Building Agents” (2024)
  • Google ADK docs (2025)

© 2025 MONTREAL.AI — Apache‑2.0

View README on GitHub